W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: Caching data returned from POST, and conditional POST

From: Brian Gaines <gaines@cpsc.ucalgary.ca>
Date: Sun, 31 Dec 1995 14:06:36 -0700
Message-Id: <v01510103ad0ca61d6721@[136.159.220.107]>
To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Larry Masinter writes
>As for HTTP/1.1, there's a cache subgroup that will be meeting, and
>perhaps they can deal with caching of other methods. (My guess is that
>the 'method' and attached data can just be added to the set of input
>parameters that might affect the output of a HTTP call, and that the
>primary difference between GET and POST is the presumption of Expires:
>for POST is 'immediate'.)

Is there any need for such a presumption since the server has 100% control
over what it sends in the Expires: field?

Netscape Navigator 2.0b4 does take note of the Expires: field, and reposts
the request. This was how I noted that it makes the request under these
circumstances a conditional POST.

We could certainly do with a very clearly defined architecture for caching
at proxy servers and browsers so that the intended usage of each controlling
field is well-defined. It should be as general as possible to
support a wide range of client-server architectures. The essence is to
provide the information through the protocol that enables servers and
browsers to coordinate effectively.

Since the designer of a web-based transaction processing system has total
control at the server end but will be using commercial browsers, it is
important in the specification to put most of the design flexibility at
the server. The browser should be a well-defined automaton responding
to information from the server. This still leaves scope for user control
at the browser if required, e.g. turning off browser interaction with the
server for "Back" if required so that a past, and possibly inconsistent,
state can be inspected as Shel suggested. However, as is the situation
with local control of typography now, most users will not be that
sophisticated and will prefer to use a system that behaves as the designer
specifies.

b.


Dr Brian R Gaines               Knowledge Science Institute
                                University of Calgary
gaines@cpsc.ucalgary.ca         Calgary, Alberta, Canada T2N 1N4
403-220-5901  Fax:403-284-4707  http://ksi.cpsc.ucalgary.ca/KSI
Received on Sunday, 31 December 1995 13:08:00 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:38 EDT